Is the wavefunction/quantum_state in electron interference experiments real?

[forkosh]

Several recent arxiv articles like
http://arxiv.org/abs/1409.6290/
reviewing the pbr theorem
http://arxiv.org/abs/1111.3328/
got me thinking about this again with respect to
straightforward two-slit interference, which I'd
thought simply and unambiguously resolved the issue
in favor of "it's real". And that resolution goes
back to (and likely before) Feynman's general audience
lectures
http://en.wikipedia.org/wiki/The_Character_of_Physical_Law
https://www.amazon.com/dp/0679601279/?tag=pfamazon01-20


You can perform a two-slit electron interference experiment,
and slowly reduce the luminosity/intensity of the source until
it's so low that you're virtually guaranteed that only one
electron at a time passes through the slits. And then accumulated
counts at the scintillation screen still exhibit interference.
So doesn't that simply mean the wavefunction/quantum_state is
ontologically "real"? If it just represented epistemic ensemble
statistics, there wouldn't be any self-interference from a
low-luminosity source.
That argument seems pretty obvious, so I guess it must be flawed, or I'd imagine it would be widely used.
So what's wrong with it?

 

[atyy]

That argument seems pretty obvious, so I guess it must be flawed, or I'd imagine it would be widely used.
So what's wrong with it?

In the PBR paper, "real" or "ontic" is used in a very specific sense. The question is if quantum mechanic is not the final theory, what sort of theories can be possible beyond quantum mechanics? Two broad classes of theories beyond quantum mechanics are considered: those in which a unique wave function is a coarse grained version of more fundamental variables, and those for which knowing the more fundamental variables does not uniquely specify the emergent wave function. In a sense, "real" is a very misleading term, and the question is: Is the wave function is in one-to-one correspondence with a specific configuration of more fundamental variables?

 

[bhobba]

You can perform a two-slit electron interference experiment,
and slowly reduce the luminosity/intensity of the source until
it's so low that you're virtually guaranteed that only one
electron at a time passes through the slits. And then accumulated
counts at the scintillation screen still exhibit interference.
So doesn't that simply mean the wavefunction/quantum_state is
ontologically "real"? If it just represented epistemic ensemble
statistics, there wouldn't be any self-interference from a
low-luminosity source.

Scratching head. When the screen registers a 'flash' that's an observation. The state tell's the probability of that. It is exactly the same as the probabilities you assign to a coin - they aren't in any sense real either.

Thanks
Bill

 

[stevendaryl]

Scratching head. When the screen registers a 'flash' that's an observation. The state tell's the probability of that. It is exactly the same as the probabilities you assign to a coin - they aren't in any sense real either.

It occurred to me that there is a kind of paradox here, even classically. On the one hand, classical probabilities are due to ignorance about the initial state, and so are not objective. On the other hand, the law of large numbers says that we can get an experimental value for the probability of a repeatable event. How can you have an experimental value for a variable, if it's not real? If two people assign different probabilities to a coin flip results, and you flip the coin often enough, then eventually one or the other will be proved wrong. So probability seems objective.

I can sort of see my way out of this paradox, and it's this: The notion of repeatability of a probabilistic event is itself subjective. For an experiment to be repeatable, you have to be able to recreate the initial state at will. But for a probabilistic experiment, the relevant notion of state reflects the ignorance of the experimenter--as far as he is concerned, it's the same state, because he is ignorant of the microscopic details that show the differences. To another person, who maybe knows more about those microscopic details, it's not the same state each time, and so he can't apply his notion of probability (which requires putting the system in the same state many times).

There is something a little unsatisfying about it, though. If I flip a coin 1 million times, and get 500 thousand heads, and 500 thousand tails, that's certainly an objective fact, independent of a person's level of ignorance. The tendency for the ratio to be 50/50 seems to be an objective fact, as well. So I'm a little confused.

 

[bhobba]

On the other hand, the law of large numbers says that we can get an experimental value for the probability of a repeatable event. How can you have an experimental value for a variable, if it's not real?

The event is assigned a probability. Its purely a semantic issue if you you consider such a thing real. I don't, and I think most wouldn't. But I am not going to argue about it because there is no right or wrong answer - semantics never do.

In the Bayesian view for example its purely a degree of confidence a human being has. In the Kolmogorov view its even more abstract than that.

The law of large numbers allows us to determine it - but that doesn't make it real like say an electric field. Another way of looking at it is you can only determine it from a large ensemble.

Thanks
Bill

 

[stevendaryl]

The event is assigned a probability. Its purely a semantic issue if you you consider such a thing real. I don't, and I think most wouldn't. But I am not going to argue about it because there is no right or wrong answer - semantics never do.

In the Bayesian view for example its purely a degree of confidence a human being has. In the Kolmogorov view its even more abstract than that.

The law of large numbers allows us to determine it - but that doesn't make it real like say an electric field. Another way of looking at it is you can only determine it from a large ensemble.

Thanks
Bill

But is it an objective property of the ensemble? By that, I mean, can different people consistently assign different values to it?

The way "ensemble" is often used, the ensemble itself doesn't literally exist, it's just a conceptual thing, so it doesn't have any objective properties. But on the other hand, sometimes the ensemble really reflects many copies of the same system.

One of the hallmarks of Bayesian statistics is that if you repeat the same experiment often enough, then everybody's posterior probabilities will converge to the same value, even when their prior probabilities were wildly different. But converging to a unique value seems to me to mean that that value is objective, in some sense.

 

[bhobba]

then everybody's posterior probabilities will converge to the same value, even when their prior probabilities were wildly different. But converging to a unique value seems to me to mean that that value is objective, in some sense.

But the key thing is it resides in a persons head.

For the Kolmogorov view its entirely abstract.

Thanks
Bill

 

[atyy]

But the key thing is it resides in a persons head.

Actually, if you look at the PBR theorem, it is about the relationship between probabilities of hidden varaibles and the wave function. The PBR argument will go through if one takes the pure Bayesian view that probabilities are in a person's head, and the hidden variables are just parameters whose existence are guaranteed by the Bayesian de Finetti representation theorem. So the PBR theorem places constraints on how one unreal thing is related to another unreal thing. There is no need to assume the reality of the hidden variables, they can be simply Bayesian parameters which are objective in the Bayesian sense that, as stevendaryl said, people with two different priors can converge to the same posterior.

 

[bohm2]

But converging to a unique value seems to me to mean that that value is objective, in some sense.

This is one of the criticisms against quantum Bayesian, in general. As Timpson points out:

The form of the argument, rather, is that there exists a deep puzzle if the quantum Bayesian is right: it will forever remain mysterious why gathering data and updating according to the rules should help us get on in life. This mystery is dispelled if one allows that subjective probabilities should track objective features of the world.

But I've seen arguments by others including by posters on here (like Ken G) that has suggested that this argument is faulty. I'm not sure what the answer is.

 

[atyy]

This is one of the criticisms against quantum Bayesian, in general. As Timpson points out:

But I've seen arguments by others including by posters on here (like Ken G) that has suggested that this argument is faulty. I'm not sure what the answer is.

It is very important to note that Quantum Bayesianism is not what stevendaryl is talking about. Bayesianism is compatible with Newtonian Mechanics and Bohmian Mechanics.

 

[bohm2]

It is very important to note that Quantum Bayesianism is not what stevendaryl is talking about. Bayesianism is compatible with Newtonian Mechanics and Bohmian Mechanics.

But isn't there a difference between Quantum Bayesianism and Bayesianism, in general? I always assumed that Quantum Bayesianism is not compatible with Bohmian Mechanics. In the latter, ψ is ontic, unlike the former.

 

[atyy]

But isn't there a difference between Quantum Bayesianism and Bayesianism, in general? I always assumed that Quantum Bayesianism is not compatible with Bohmian Mechanics. In the latter, ψ is ontic, unlike the former.

Quantum Bayesianism and Bayesianism are different. But it's confusing that you quoted stevendaryl who was only talking about Bayesianism, and then suddenly saying that what he said is a criticism of Quantum Bayesianism.

 

[forkosh]

Scratching head. When the screen registers a 'flash' that's an observation. The state tell's the probability of that. It is exactly the same as the probabilities you assign to a coin - they aren't in any sense real either. Bill

That (what you said) is the epistemic interpretation. But it doesn't explain self-interference. If the source luminosity is so low that you're guaranteed just one electron at a time through the slits, then the still-observed interference means those individual electrons must each be "going through both slits" in some kind of way. That's the ontic interpretation, i.e., the probability tells you "how much of each individual electron" goes through one slit, and how much through the other. That is, the electron propagates like a wave, thus going through both slits, but interacts like a particle, thus registering discrete measurement events at the screen.

See, for example, http://hyperphysics.phy-astr.gsu.edu/hbase/phyopt/slits.html if you're not familiar with two-slit interference. Something has to be going through both slits for interference to occur. Interpreting the probability as statistics won't account for that.

 

[bhobba]

That (what you said) is the epistemic interpretation. But it doesn't explain self-interference. If the source luminosity is so low that you're guaranteed just one electron at a time through the slits, then the still-observed interference means those individual electrons must each be "going through both slits" in some kind of way. That's the ontic interpretation, i.e., the probability tells you "how much of each individual electron" goes through one slit, and how much through the other. That is, the electron propagates like a wave, thus going through both slits, but interacts like a particle, thus registering discrete measurement events at the screen.

I know it - and here is the QM explanation:
http://arxiv.org/ftp/quant-ph/papers/0703/0703126.pdf

Its equally valid regardless of interpretation. nor does anything have to be going through the slits.

Thanks
Bill

 

[forkosh]

I know it - and here is the QM explanation:
http://arxiv.org/ftp/quant-ph/papers/0703/0703126.pdf
Its equally valid regardless of interpretation. nor does anything have to be going through the slits.
Bill

Okay, so that's more quantitative, but still in agreement with what I qualitatively said. And in disagreement with your earlier epistemic remark about the interpretation of slit probabilities as (your words) "same as the probabilities you assign to a coin" -- re-read your own reference to see that your epistemic (probabilities are just statistics) interpretation can't be right.

Moreover, the slits are necessary to exhibit self-interference. The very first sentence of your own reference's abstract says "...a particle passing through a system of slits." And then the very first sentence of the article itself says, "The double-slit experiment is the archetypical system used to demonstrate quantum mechanical behavior." So you can't cite that reference -- your own reference -- and say "...nor does anything have to be going through the slits." That's patently contradicting the reference you're citing. And it's simply not right.

And that leaves my original question unaddressed: the easily observable two-slit self-interference seems to simply and unambiguously confirm the ontic interpretation of the quantum state. Indeed, I've given you the Feynman lecture link -- you can watch Feynman make that argument for yourself. Although at the time of his 1964 lecture, the terms ontic and epsitemic weren't used much, so you have to read between the lines a little and re-word his conclusion.

But I don't see Feynman's argument used much in the contemporary debate. So why not, or what's wrong with it, or with my interpretation of it?

 

[vanhees71]

Well, Feynman wouldn't discuss philosophy. For him it's "cargo-cult science". Sometimes I have the strong feeling that he was right with this assessment of philosophy, particularly when I read philosopher's writings on quantum theory. There are exceptions, of course.

Anyway. I also tend to an epistemic interpretation of the quantum mechanical state. The problem with ontic interpretations is always, what's dubbed "collapse", which contradicts explicitly Einstein causality in saying that a delocaliced electron becomes localized instantly when being registered. So, if you say the Schrödinger wave function (let's discuss the nonrelativistic limit for simplicity for the moment) "is the electron" in some way, this implies that a local interaction of the electron with the detector leads to its immediate collapse in far-away regions, which contradicts Einstein causality.

Another good argument against such a ontic interpretation is the fact that strictly speaking in relativistic quantum theory there's no such thing as a single-particle wave function, since you cannot localize a particle without necessarily creating new particles, i.e., the single-particle wave-function interpretation doesn't hold anymore. This is a pretty general argument by Bohr and Rosenfeld, which is even independent of the realization of relativistic QT as local microcausal QFT, which are however the only (very!) successful relativistic QT for real-world physics (aka the "Standard Model of Elementary Particles").

 

[bhobba]

Okay, so that's more quantitative, but still in agreement with what I qualitatively said. And in disagreement with your earlier epistemic remark about the interpretation of slit probabilities as (your words) "same as the probabilities you assign to a coin" -- re-read your own reference to see that your epistemic (probabilities are just statistics) interpretation can't be right.

It is? Can you please explain why without some kind of hand-wavy argument such as something must be going through both slits. I am looking for a clear concise mathematical argument such as the paper I linked which explains it from the simple fact the slits are a position observation.

And if you can't do that you may have to just consider the fact it may not be real in a physical sense.

As for the wording of some parts of the paper its simple context stuff you find all over the place in physics and applied math - people are not always 100% careful in what's said when dealing with subtle issues. Even though nothing necessarily has to be going through the slits wording like that is often resorted to. Remember physics is not philosophy where semantics is paramount - physicists are much more pragmatic about such things.

QM has been a mystery for many decades - it can not be waved away with simple platitudes such as something must be going through the slits - if it was that easy the issues would have been resolved ages ago.

Thanks
Bill

 

[bhobba]

Well, Feynman wouldn't discuss philosophy. For him it's "cargo-cult science". Sometimes I have the strong feeling that he was right with this assessment of philosophy, particularly when I read philosopher's writings on quantum theory. There are exceptions, of course.

On this, as on many other issues, he penetrated to heart of the matter.

Thanks
Bill